Here's something I've always wondered. I use Monster Power conditioners, a 1600 and 2500. I paid next to nothing for them and use them mainly for the surge protection. I did notice a slight improvement in sound quality from my equipment but nothing earth shattering. If I were to add a mains filter to each device after the power conditioner would it make any positive difference? Or is the filter in the power conditioners do all the smoothing that can be done so to speak?
That makes sense to me. I would be filtering DIY sources and a preamp. All my amps are low power headphone based so they're not heavy loads. I think I'll start adding them at least on the source and preamp side. That makes sense to me
One good approach to minimizing noise is to plug every component into the same outlet strip.
Also try reversing any plugs (one at a time) that have only two prongs. But don't use cheater plugs.
Various electronic devices are notorious for making noise, so try turning some off if you have any stubborn noise.
Also try reversing any plugs (one at a time) that have only two prongs. But don't use cheater plugs.
Various electronic devices are notorious for making noise, so try turning some off if you have any stubborn noise.
Alas some powerline filters do more harm than good. They present a high impedance to noise, which might be good at keeping noise out, but they can also keep it in!
The low impedance of the mains can actually work as a noise sink. Put a high impedance filter in line, and the noise has nowhere to go. You get more noise at the device power plug than with no filter.
Yes, I have measured this, out to several MHz.
The low impedance of the mains can actually work as a noise sink. Put a high impedance filter in line, and the noise has nowhere to go. You get more noise at the device power plug than with no filter.
Yes, I have measured this, out to several MHz.
That's interesting and sort of what I was afraid of. My thinking was a filter being filtered could cause it's own effects just wasn't sure if it would be good or bad. I've never filtered inside a source or amp so that's why I questioned it. I built a new house in FL where the A/C compressor, air handler, and fridge run more often than not so I was thinking of things to suppress the bad stuff that comes from that. Plus lightning strikes and storms as well. We already had a few homes in our community get hit. I have a surge protector on the power meter and run surge protectors on everything else on top of that so we haven't had any issues so far.
I spent a lot of time in Florida and so well understand the fridge, AC, lightening and other electrical noises. 
Mains filtering of some sort seems like a good idea, I would just say "be careful" and measure if you can.
Years ago when I posted my line filter finding on another forum, my results were met with a great deal of hostility from filter designers. They claimed that no good filter would ever behave that way - I could only counter by saying that the results were clear on this large, expensive filter built expressly for audio use.
Surely there must be good and bad filters, but how does one know, without actually measuring the noise on both sides of the filter? I was lucky to be able to that.
Mains filtering of some sort seems like a good idea, I would just say "be careful" and measure if you can.Years ago when I posted my line filter finding on another forum, my results were met with a great deal of hostility from filter designers. They claimed that no good filter would ever behave that way - I could only counter by saying that the results were clear on this large, expensive filter built expressly for audio use.
Surely there must be good and bad filters, but how does one know, without actually measuring the noise on both sides of the filter? I was lucky to be able to that.
I don't have a way to measure but I will let my ears be my guide. I think what I'll do is make a small box with just the filter inside with an IEC inlet on one side and an AC outlet on the other. Or just get a metal junction box, wire up some SJ cable with a plug, connect it to the filter and then the filter to an outlet with a faceplate on the front of the junction box. That'll let me plug and play so to speak. If I hear a difference I'll add it to all my source and preamp builds.
I'd buy an inexpensive scope off Aliexpress but I have no idea how to use it so for now it'll be my ears
I'd buy an inexpensive scope off Aliexpress but I have no idea how to use it so for now it'll be my ears
Such a scope by itself would be useless and dangerous, since it is not isolated.
To probe the AC line, an isolated differential voltage probe is required.
https://siglentna.com/products/accessories/probes/differential-high-voltage-probes/
Or else one of the few scopes that have individually isolated inputs, which can use conventional probes.
https://www.fluke.com/en-us/product/electrical-testing/portable-oscilloscopes/190-series-iii
To probe the AC line, an isolated differential voltage probe is required.
https://siglentna.com/products/accessories/probes/differential-high-voltage-probes/
Or else one of the few scopes that have individually isolated inputs, which can use conventional probes.
https://www.fluke.com/en-us/product/electrical-testing/portable-oscilloscopes/190-series-iii
I was thinking of measuring for noise on the output of the source or preamp. But definitely thanks for the warning!
You could wire the specific for-audio power outlet(s) back to the power meter box via separate feeds (to individual C/Bs); use individual ground wire (NOT daisy chained). Just specify fairly thick multi-strand (litz) ground wire (NOT solid core). I suggest having 2 outlets wired in such a way - one will be for the noisy devices powered by the SMPSs (like TV/PVR/noisy SMPS power packs...), and the other only for the true analog audio equipment (analog pre-amps/amps). This will make a huge difference to the overall sound quality reproduction. Electricians usually use cheap stuff, like solid core neutral and phase (which is okay), with another solid core for ground (NOT okay... does not do much at high frequencies... can't couple that rubbish to ground (well, to neutral at the power distribution box).... add a separate thick multi-strand wire. This will still be fully compliant (DC grounding) with the wiring standards. Here's the wire I'm talking about - see that ground lead... that's the one to avoid:
As for the mains RF filters, I tried many and learned to avoid them - for the exact same reason as Pano explained in post #5, so no need to repeat what he's written already.... I separate my analog audio gear from my SMPS-powered gear. In fact, I reduced the SMPS-powered gear to only a TV. Everything else uses linear power supplies (streamer as well).
Last edited:
IMHO your measurements agree with what transmission line theory would predict: Wherever you have a change in impedance, some energy is reflected back to the sender (measurable as the reflection coefficient, which is related to the relative impedances). I always wondered if it makes sense to send something back to sender, which obviously cannot handle it properly. Otherwise it wouldnt send it, correct?
I think a possible escape could be insertion of ohmic losses, which transform this "filtered" energy into heat. For example, in an AC filter placed after the inverter of my photovoltaics I connected the join of the Y capacitors not directly to ground but with a 10R resistor. This, of course, doesnt eliminate all the noise above 50Hz, but at least some of it. A chain of RC shunts might therefore improve the elimination of high-frequency noise better than a series of lossless CL filters (as shown and measured by @Pano ).
The point being that pollution is/should be directed to PE. That is a low ohmic shunt path. Also that filtering works both ways, it is not only pollution coming in but also pollution going out and this counts especially for modern equipment, computers etc. with SMPS. RF can be filtered out and it will be beneficial to much linear stuff where RF just wanders through normally. Even the best of old linear regulating technology is not resistant against today's pollution without taking measures. It speaks for itself to avoid Aliexpress and such stuff but please choose quality branded industrial shielded filters (preferably IEC inlets with builtin fusing and filtering) that do work as intended.
Two prong situations are incomparable really.
Anything that involves DC-AC inverters pollutes and needs filtering of sorts not only to filter out RF but also to improve sinusoid waveforms. A good example are the older Mastervolt Soladin 600 solar inverters that used the solar panels (unintended) as antenna and blocked telecom. Luckily solar is implemented everywhere leading to noise ridden mains voltage, aerial RF and also higher mains voltage. High power online IGBT based UPS systems also can not do without filtering.
https://www.briandorey.com/post/mastervolt-soladin-rf-filtering
Two prong situations are incomparable really.
Anything that involves DC-AC inverters pollutes and needs filtering of sorts not only to filter out RF but also to improve sinusoid waveforms. A good example are the older Mastervolt Soladin 600 solar inverters that used the solar panels (unintended) as antenna and blocked telecom. Luckily solar is implemented everywhere leading to noise ridden mains voltage, aerial RF and also higher mains voltage. High power online IGBT based UPS systems also can not do without filtering.
https://www.briandorey.com/post/mastervolt-soladin-rf-filtering
Last edited:
I am always surprised by people saying that a "linear" power supply (switching at 100/120 Hz) sounds better than a "linear" power supply switching at 50 kHz or more. It is far easier to filter 50 kHz than 100/120 Hz. There exist no linear power except for batteries. You can just as easily regulate any power supply to behave linearly like batteries provided you have the necessary skills.
I avoid using huge expensive transformers and capacitors in any equipment these days. You can filter a switch mode and regulate it far better than a "linear"power supply much more efficiently and cost effectively.
I avoid using huge expensive transformers and capacitors in any equipment these days. You can filter a switch mode and regulate it far better than a "linear"power supply much more efficiently and cost effectively.
Figure 13-1 in Ott (l.c.) shows that minimum impedances in a 115V power line range from about 3 (0.1MHz) to 30 Ohms at 30MHz, maximum values go up to 300-500 Ohm. Is this still a "low ohmic shunt path"?
In another publication (in German) Weitnauer describes, how the usage of different kinds of power cables in a building are very likely responsible for peaking phenomena of impedances in a building. Capacitive coupling between cables in combination with the cable inductance makes resonance circuits, which resonate in the MHz range. Depending on where the measurements were made, impedances exceeded 100Ohm (at 300 kHz). A nice example, why anti-noise measures may be effective in one location, but less in another one.
All in all a highly complicated field, where without proper measurement capabilities only trial and error may succeed.
Source of figure: Nicholson & Malack, IEEE Transactions on Electromagnetic Compatibility, May 1973).
Weitnauer, Bulletin SEV/VSE 23/02
Once measured and established, you must fix it. Not so easy working in high current high voltage changing load environment.
Tell me. It has become more difficult with complex and capacitive loads.
The past years I have had generators stopping and UPSes overheating when massive amounts of neon lighting were replaced for LED lighting. Sometimes unexpected behavior like high power SMPSes operated with light loads and/or power saving enabled at the load causing odd issues like a “wandering” cosinus phi. Also a rising necessity to improve mains voltage quality and of course cosinus phi improvement. Power analysers used to be a niche tool but today less so with DC-AC inverters being main grid maintainers in areas. In the near future DC will likely be used for power distribution. That is good, new challenges.
In industrial power generation/conversion devices there is no debate, they all have filtering.
The past years I have had generators stopping and UPSes overheating when massive amounts of neon lighting were replaced for LED lighting. Sometimes unexpected behavior like high power SMPSes operated with light loads and/or power saving enabled at the load causing odd issues like a “wandering” cosinus phi. Also a rising necessity to improve mains voltage quality and of course cosinus phi improvement. Power analysers used to be a niche tool but today less so with DC-AC inverters being main grid maintainers in areas. In the near future DC will likely be used for power distribution. That is good, new challenges.
In industrial power generation/conversion devices there is no debate, they all have filtering.
Last edited:
I am curious about that, because I wonder if the Power Earth is also rather dirty (compared to what?) A number of times I have poked probes into the dirt about 3 meters apart. What did I hear? Buzzzzzzzzz. A whole lot of mains noise, mostly H2 and H3. It's quite possible my tests meant nothing, but they did worry me. 😉
This is something that I had wanted to do when I was setting up my shop/office. I ended up not doing it as I was unsure where I was going to place things. Now that I have a goo idea where everything will be I could have the electrician make one run configured like this.
So the ground wire would be 12 AWG stranded, but I run it outside of the 12/2 jacket from end to end? Is that right?
I think it was Ott who said in one of his papers that the term "ground" or "earth" is probably one of the most misunderstood terms. It would be great if "ground" would absorb all the HF noise we produce now amply with our "earth-saving" (pun intended) SMPS wall-marts. With a current probe around the PE it should be possible to measure its pollution level. For some measurements of "ground" noise see a talk by Jim Brown (jim@audiosystemsgroup.com) "RFI in Audio Systems Pin 1 problems, poor shielding, and poor input/output filtering"